-
Profi ling of Polycyclic Aromatic Hydrocarbons in Crude Oil with
the Agilent 1290 Infi nity 2D-LC Solution
Application Note
AuthorsGerd Vanhoenacker, Frank David, and Pat
SandraResearch Institute for ChromatographyKennedypark 26B-8500
KortrijkBelgium
Udo HuberAgilent Technologies, Inc.Waldbronn, Germany
Energy and Chemicals
AbstractThe Agilent 1290 Infi nity 2D-LC Solution was used to
profi le the polyaromatic hydrocarbon (PAH) fraction from mineral
oil using comprehensive two-dimensional liquid chromatography
(LCxLC). The complexity of this fraction, consisting of
nonsubstituted PAHs, alkyl-substituted PAHs, and heterocyclic PAHs,
largely exceeds the peak capacity of a one-dimensional LC
separation.
A combination of a cyanopropyl column in the fi rst dimension
and a dedicated PAH column in the second dimension provided good
orthogonality, resulting in higher peak capacity. Detection was
performed by parallel diode-array and fl uorescence detection. This
Application Note shows the potential of 2D-LC for profi ling the
polyaromatic fraction of mineral oils.
-
2
InstrumentationAn Agilent 1290 Infi nity 2D-LC Solution with the
following confi guration was used for the experiments.
• Agilent 1290 Infi nity Binary Pump, for fi rst dimension
(G4220A)
• Agilent 1290 Infi nity Binary Pump, for second dimension
(G4220A)
• Agilent 1290 Infi nity Autosampler (G4226A)
• Agilent 1290 Infi nity Thermostat (G1330A)
• Agilent 1290 Infi nity Thermostatted Column Compartment
(G1316C)
• Agilent 1290 Infi nity Diode Array Detector with standard fl
ow cell (G4212A)
• Agilent 1260 Infi nity Fluorescence Detector (G1321B)
• Agilent 1290 Infi nity Valve Drive (G1170A)
• Agilent 1290 Infi nity 2-position/4-port duo-valve for 2D-LC
(G4236A)
Software• Agilent OpenLAB CDS
ChemStation Edition software, version C.01.07, with 1290 Infi
nity 2D-LC software, version A.01.02
• GC Image LCxLC Edition software for 2D-LC data analysis (GC
Image, LLC., Lincoln, NE, USA)
The aromatic and polycyclic aromatic fraction contains
nonsubstituted two to six-ring PAHs, alkyl-substituted PAHs,
heterocyclic PAHs (for example, dibenzothiophene),
alkyl-substituted heterocyclic PAHs and, possibly, more polar
derivatives such as hydroxy-PAHs, amino-PAHs, and nitro-PAHs. Due
to this high complexity, high resolution separation techniques are
needed. MOSH and MOAH fractions are commonly analyzed by GCxGC and
GCxGC/MS, but comprehensive LCxLC can be considered as an excellent
complementary technique, especially since the high molecular weight
PAHs (six rings) can be easily analyzed, and selective detection by
fl uorescence is very sensitive.
The combination of two separation modes, such as a ring number
separation with a hydrophobicity separation, can be particularly
useful for profi ling the aromatic fraction in oils. This
Application Note, illustrates the LCxLC approach using the Agilent
1290 Infi nity 2D-LC Solution.
ExperimentalSamples and sample preparationThe standard solution
containing 16 PAHs in acetone/benzene at a concentration of
2 mg/mL each (PAH Mix 25, Dr. Ehrenstorfer, Augsburg, Germany)
was diluted to 10 µg/mL in acetone.
From a crude oil sample, the polyaromatic fraction was isolated
using a liquid-liquid partitioning between hexane and
nitromethane.
A sample of 100 mg of crude oil was dissolved in 5 mL
hexane. After dissolution, 5 mL of nitromethane was added and
a liquid-liquid extraction was performed. The upper hexane fraction
contained the saturated hydrocarbons bulk fraction. The lower
nitromethane layer, containing the more polar aromatic fraction,
was collected for analysis. For a crude oil sample, the aromatic
fraction is typically 5 to 30 % of the total sample1.
IntroductionPolycyclic aromatic hydrocarbons (PAHs) are
well-known contaminants in the environment and in food samples.
PAHs mostly originate from natural and anthropogenic combustion
processes. To date, most analytical methods for the trace-level
analysis of PAHs in environmental samples (soil, sediment, water,
air) and in food samples (mostly fatty foods), focus on a selected
number (typically 16) of nonsubstituted polycyclic aromatic
hydrocarbons, such as fl uoranthene, chrysene, benzo(a)pyrene, and
benzofl uoranthenes. Analytical methods are based on GC/MS
(including single quadrupole and triple quadrupole MS) or on HPLC
in combination with diode-array detection (DAD) or fl uorescence
detection (FLD).
Carcinogenity of certain PAHs has been unequivocally
demonstrated and, similar to polychlorinated dioxins and furans
(PCDDs/PCDFs) and polychlorinated biphenyls (PCBs), toxicity
equivalent factors (TEF) are used to measure the total
contamination of a sample by PAHs.
More recently, concerns have been raised regarding the toxicity
of alkyl-substituted PAHs. Indeed, in petroleum products such as
diesel, mineral oils, and crude oils, the contribution of
substituted PAHs to the total (poly)aromatic fraction is much
larger than the contribution of the nonalkylated PAHs that are
typically analyzed. This fact is, for instance, recognized by EFSA
in their Scientifi c Opinion on Mineral Oil Hydrocarbons in Food1.
For the analysis of mineral oil in food and in packaging material,
GC-FID methods are used after a preseparation of the saturated
hydrocarbons (mineral oil saturated hydrocarbons, MOSH) fraction
from the aromatic fraction (mineral oil aromatic hydrocarbons,
MOAH) using solid phase extraction or normal phase HPLC.
While measuring the MOSH fraction is well documented, good
characterization of the aromatic fraction is still lacking. This is
contradictory with the fact that toxicity of the aromatic fraction
is substantially higher than that of the saturated hydrocarbon
fraction.
-
3
MethodFirst dimensionColumn Agilent ZORBAX SB-CN, 2.1 × 150 mm,
5 µm (p/n 883700-905)Solvent A WaterSolvent B MethanolFlow rate 100
µL/minGradient 40 %B at 0 minutes
100 %B at 80 minutes100 %B at 85 minutes
Posttime 10 minutes at 40 %BColumn temperature 40 °CSecond
dimensionColumn Agilent ZORBAX RRHD Eclipse PAH, 3.0 × 50 mm, 1.8
µm (p/n 959757-318)Solvent A WaterSolvent B AcetonitrileFlow rate 2
mL/minIdle fl ow rate 0.3 mL/minInitial gradient 50 to 70 %B from 0
to 0.35 minutes
70 %B from 0.35 to 0.40 minutes 50 %B at 0.41 minutes
Gradient modulation 50 %B at 0 minutes to 100 %B at 70 minutes70
%B at 0.35 minutes to 100 %B at 55 minutes
Column temperature 40 °CModulationModulation on 7 to 85
minutesLoops Two 60-µL loops, cocurrent confi gurationModulation
time 0.50 minutesInjectiona
Volume 1 µL (injection program, mixed with 1-µL water plug
)Needle wash 5 seconds fl ush port (methanol/acetone)Detection
DADb
Wavelength Signal 220/10 nm Data rate 80 HzDetection FLDb
Wavelength Multi-emission modeSignal A: Ex 260 nm/Em 350
nmSignal B: Ex 260 nm/Em 430 nmSignal C: Ex 260 nm/Em 500 nm
Data rate 37.04 HzPMT Gain 7
a The samples were injected together with a water plug to avoid
peak broadening/splitting due to the strong injection solvent.
b A zero-dead volume T-piece was installed at the outlet of the
second dimension column to split the fl ow between the DAD and FLD.
Red 0.12-mm PEEK tubing was used to connect to the detectors. The
tubing length from the T-piece to the FLD was twice as long
compared to the tubing going to the DAD, resulting in a DAD/FLD
split ratio of about 2:1.
-
4
Results and DiscussionThe LC analysis of PAHs is generally
performed using dedicated PAH columns and DAD or FLD detection.
Using this column chemistry with water/acetonitrile gradients
results in complete separation of the 16 most important
nonsubstituted PAHs. When this sample is analyzed on other reversed
phase systems (other columns, other mobile phases), some coelution
typically does occur, but the majority of analytes are separated.
The analyses of PAHs in more complex samples (number of PAHs or
complexity of the matrix) are signifi cantly more challenging and
require more chromatographic selectivity and separation power and,
if possible, better detection selectivity. A comparison of the
one-dimensional analysis of the standard mixture of 16 PAHs and the
crude oil extract on the fi rst dimension ZORBAX SB-CN column is
shown in Figure 1 (chromatographic conditions are different from fi
nal LCxLC conditions). It is clear that the complexity of the real
sample is far too high for one-dimensional separations. The
complexity of the sample originates from the fact that the crude
oil contains substituted PAHs next to the nonsubstituted PAHs, and
that heterocyclic PAHs are also present. The substituted PAHs are
mainly alkylated PAHs, and make up a large group of solutes taking
into account variations in substitution degree, alkyl chain length,
branching, and substitute position. All these result in extreme
complexity and no chromatographic technique is available that would
achieve full resolution of all possible individual compounds. For
the characterization of the polyaromatic fraction, it is important
to obtain information on the number of aromatic rings, relative
substitution degree, and on the possible presence of more polar
PAHs.
On the selected fi rst dimension SB-CN column, an interesting
separation is obtained. The elution order for the 16 PAHs
under the applied conditions differs signifi cantly from the normal
elution order in a classical RP-LC PAH analysis (as in Table 1).
This opens perspectives for orthogonality in the 2D-LC setup.
T able 1. Peak identities of the 16 PAHs (peak numbers are
assigned according to the expected elution order on a PAH
column).
Peak no. Compound1 Naphthalene2 Acenaphthylene3 Acenaphthene4
Fluorene5 Phenantrene6 Anthracene7 Fluoranthene8 Pyrene9
Benzo(a)anthracene10 Chrysene11 Benzo(b)fl uoranthene12 Benzo(k)fl
uoranthene13 Benzo(a)pyrene14 Dibenzo(ah)anthracene15
Benzo(ghi)perylene16 Indeno(1,2,3-cd)pyrene
Time (min)0 5 10 15 20 25
Time (min)0 5 10 15 20 25
mAU
050
100150200250300350
mAU
050
100150200250300350
1
23
4 5 6 8 910 12,13 15
PAH mix 16
A
B
Crude oil extract
Acetone (solvent)
Benzene(solvent)
16
147 11
F igure 1. Comparison of a one-dimensional analysis of the PAH
standard mix and sample extract. Column: Agilent ZORBAX SB-CN, 2.1
× 150 mm, 5 µm, Flow rate: 0.3 mL/min, Gradient: 40–100 %
methanol in water from 0–40 minutes. Peak identities: see Table
1.
-
5
Adapted chromatographic conditions and switching to UHPLC
equipment could improve the resolution for the crude oil extract,
but to drastically increase peak capacity, comprehensive 2D-LC will
be more effective. A combination was made of the ZORBAX SB-CN
column using a water/methanol mobile phase in the fi rst dimension,
with a classic PAH analysis setup in the second dimension (Eclipse
PAH column with a water/acetonitrile mobile phase). On the
dedicated PAH column, better separation was obtained within a group
with equal ring number (for example, benzofl uoranthene
isomers).
The LCxLC contour plots obtained with DAD using the
SB-CN/Eclipse PAH combination is shown in Figure 2B, and can
be compared to Figure 2A, showing the 1D separation on the SB-CN
column. Better separation was obtained for phenanthrene/anthracene,
benzo(a)anthracene/chrysene, and benzofl uoranthenes. Only
benzo(ghi)perylene and indeno(1,2,3-cd)pyrene were still not
completely separated. Next, the aromatic fraction of the mineral
oil was analyzed using the same conditions. The LCxLC contour plot
is shown in Figure 2C.
Fig ure 2. Comparison of a 1D-LC run of the standard mix (A),
LCxLC run of the standard mix (B), and LCxLC run of the sample
extract (C). Signal: DAD, 220 nm.
A
Time (min)10 20 30 40 50 60
mAU
0
20
40
60
80
100
120
140
1601
2
3
4
4
5
5
5
6
6
6
8
8
8
9
9
9
10
10
13/1215
16147
7
7
11
11
15 16
1412
13
B
C
1
1
23
3
4
-
6
Figure 3, Figure 4, and Figure 5 show the LCxLC results from FLD
using different emission wavelengths. Some additional series of
compounds (probably substituted PAHs) are clearly detected. As an
example, there is a series of compounds that elutes prior to the
nonsubstituted PAHs on the second dimension column. This is clearly
visible with FLD. There is also a considerable number of compounds
that are more or less scattered around the PAHs present in the
standard mix. Identifi cation of the additional compounds in the
mineral oil extract will require further investigation with, for
example, hyphenation to MS using atmospheric pressure
photoionization (APPI), but from their relative elution pattern it
can be predicted that these are alkyl-substituted PAHs. These
results clearly illustrate the high complexity of the PAH fraction
of crude oil.
Although many compounds remain unidentifi ed, the results
clearly demonstrate the potential of the Agilent 1290 Infi
nity 2D-LC Solution for this type of analysis.
A
Time (min)
LU
B
C
10 20 30 40 50 60
0
2
4
6
8
10
12
14
16
1
1
1
3
3
4
4
4
5
5
5
8
8
8
10
10
76
Figure 3. Comparison of a 1D-LC run of the standard mix (A),
LCxLC run of the standard mix (B), and LCxLC run of the sample
extract (C). Signal: FLD, Ex 260 nm/Em 350 nm.
-
7
ConclusionThis Application Note shows the potential of the
Agilent 1290 Infi nity 2D-LC Solution for profi ling PAHs in the
aromatic fraction of mineral oils using parallel diode-array and fl
uorescence detection. The method is useful for profi ling PAHs in
crude oils, bitumen, and other mineral oils. The combination of the
presented 2D-LC method with MS using APPI ionization should
facilitate further structure elucidation of the detected PAH
compounds.
Reference1. EFSA Journal 2012; 10(6):2704,
European Food Safety Authority
http://www.efsa.europa.eu/fr/search/doc/2704.pdf
Figure 4. Comparison of a 1D-LC run of the standard mix (A),
LCxLC run of the standard mix (B), and LCxLC run of the sample
extract (C). Signal: FLD, Ex 260 nm/Em 430 nm.
A
Time (min)
LU
B
C
10 20 30 40 50 60
0
20
40
60
80
5
5
5
6
6
6
8
8
8
9
9
9
10
10
10
15
15
14
14
7
7
13/12
11
11
11
12
12
13
13
-
www.agilent.com/chem
This information is subject to change without notice.
© Agilent Technologies, Inc., 2015Published in the USA, March 1,
20155991-5619EN
Figure 5. Comparison of a 1D-LC run of the standard mix (A),
LCxLC run of the standard mix (B), and LCxLC run of the sample
extract (C). Signal: FLD, Ex 260 nm/Em 500 nm.
10 20 30 40 50 600
2
4
6
8
10
12
6 910
15
167
13/12
11A
Time (min)
LU
B
C
6
6 7
9
9
16
7
11
11
12
12
13